A frugal microfluidic pump

Abstract

Manipulation of fluid flow is paramount for microfluidic device operation. Conventional microfluidic pumps are often expensive, bulky, complicated, and not amenable in limited resource settings. Here, we introduce a Fully self-sufficient, RobUst, Gravity-Assisted, Low-cost (FRUGAL) microfluidic pump. The pump consists of a syringe, a syringe holder and loading masses. The system is easy to assemble, inexpensive, portable, and electrical power-free. Inside the syringe, the fluid is driven by the pressure from the weight of the loading masses. Therefore, during operation, the exerted pressure is dynamically controllable and stable for hours, features that are useful for optimization of microfluidics assays and dynamic temporal studies. We demonstrate the application of this system to control the formation of water-in-oil droplet emulsion. Benefitting from its simplicity and versatility, the frugal microfluidic pump will enable global adoption of microfluidic technology in chemistry and biomedical applications, especially in limited resource environments.

Graphical Abstract

A frugal microfluidic pump is a fluid actuation system that is Fully self-sufficient, RobUst, Gravity-Assisted, Low-cost (FRUGAL). The pump can generate pressure for versatile microfluidic applications.

Introduction

Miniaturization of analytical devices in microfluidics and lab-on-chip technologies holds potentials for mitigating global health care challenges.¹ In particular, microfluidics devices are often deemed powerful to facilitate sample preparation for point-of-care diagnostics.^2,3 The microfluidic device is also an emerging tool adopted in cell biology and biomedical science research as it offers high-throughput sample processing, efficient handling of fluids and culture media, and precision measurement of microscale phenomenon.^4-6 These applications include single cell droplet sequencing,⁷ high-throughput single cell deformability assays,⁸ and organ on chip models.⁹

Fluid sample manipulation is the initial step in microfluidic device operation. Passive effects, such as capillary and gravitational forces, are often harnessed to flow fluids into microfluidics.^10-12 Passive methods rely on the geometry of the microfluidic device making it difficult to control the fluid flow. In addition, miniaturized pumps can be integrated in the microfluidic chip by using mechanical,^13,14 electrical,¹⁵ magnetic,¹⁶ or acoustic actuation.¹⁷ Unfortunately, integrated micropumps have limited performance as well as small volume of sample fluid.¹⁸ For continuous flow microfluidic devices, common fluid actuation strategy relies on external pumping system that are bulky, relatively expensive and requires electrical power. Therefore, a conventional commercial microfluidics pump is inaccessible for use in low resource settings and under field conditions because it is expensive and requires electricity.

Recently, there have been some alternatives for inexpensive external pumping system. Thurgood et al. implemented a pressurized balloon to control fluid flow in microfluidics.¹⁹ The balloon can withstand up to 5 kPa of inflation pressure. To flow fluid through the microfluidic channels, the user squeezes the balloon, providing a low-pressure exertion system that is difficult to replicate. To increase the operational pressure, they added layers of stocking fibre so the balloon can exert 25 kPa of pressure.²⁰ Another system is a plug-and-play microfluidic pump, reported by Zhang et al., which consists of a spring-loaded cartridge to drive fluidic pressure of up to 70 kPa.²¹ The plug-and-play pump is simple to use but it is more complicated to fabricate and integrate with microfluidics. The use of 3D-printed peristaltic pumps has been demonstrated by Behrens et al. as well as Ching et al. to control fluid flow in microfluidics.^22,23 The pumps can be assembled using off-the-shelf parts, but still requires a power source to run the gearbox. Combining 3D printing and acoustofluidics, Ozcelik and Aslan reported a fluid pump with a limited flow rate of 12 μL/min.²⁴ A simpler option to drive fluids into microfluidic is to use handheld syringe.²⁵ The pressure is set by holding the syringe plunger with a wire. However, the drawback of the operation is the inability to dynamically control the exerted pressure. Overall, there is a lack of external pumping technology that is low-cost, field-operable, and versatile.

We address the current limitations to control fluid flow in microfluidics by a Fully self-sufficient, RobUst, Gravity-Assisted, Low-cost (FRUGAL) microfluidic pump. The frugal microfluidic pump system comprises of a syringe, a syringe holder, and loading masses. This simple system, in which loading mass is utilized to actuate the syringe, has never been reported before. The system is easy to fabricate, requires no expertise in electronics or machining. The fabrication cost is minuscule (< US$ 2) and can be adjusted by choosing inexpensive material for the syringe holder. The system is small, portable, and requires no electricity for use, therefore making it suitable for in situ field study using microfluidics. The frugal microfluidic pump can generate an operating pressure of more than 700 kPa (100 Psi), outperforming by almost 2 orders of magnitude in comparison to the previously demonstrated low-cost systems.¹⁹ Moreover, the exerted pressure is also stable for a long period of time and can be dynamically controlled. These characteristics demonstrate the applicability of the system for many longitudinal assays in biology and biomedical research. We also show the ability of the pump to precisely control droplet size in a microfluidic droplet generator. The overall features of the frugal microfluidic pumps are expected to increase the adoption of microfluidic technology by non-experts as well as implementation in resource-limited settings.

Experimental

Working principles

The proposed frugal microfluidic pump is shown in Fig. 1. The input of the system is the loading mass. For characterization purposes, a calibrated weight balance is used as the loading mass. For ease of portability and to reduce cost, a plastic measuring cup can be used instead. As the vertical actuation of the syringe relies on the weight of the mass, it is important to keep the frugal microfluidic pump on a level surface. Following Pascal’s law, the pressure inside a confined incompressible fluid is equal. The exerted pressure results from the weight of the loading mass on the inner cross section of the syringe. Therefore,

$$P_{exert}=P_{load}=\frac{w_{load}}{A_{syringe}}=\frac{m_{load}\times g}{\frac{1}{4}\times \pi \times {d_{syringe}}^2}$$ (1)

where m_load is the loading mass, g is gravitational acceleration (9.82 m/s²), and d_syringe is the inner diameter of the syringe. According to this equation, two main factors that govern the exerted pressure are the loading masses and the syringe diameter (that is related to the syringe size).

Figure 1.

Concept illustration and photograph of the assembled frugal microfluidic pump.

Frugal microfluidic pump fabrication

The system contains a syringe, a syringe holder, and loading mass. The syringe holder consists of two segments: a static syringe plate to hold the syringe in place and a mass plate to apply the loading mass to the plunger. Each of these components were constructed from 6.35 mm (1/4^th inch) thick acrylic sheet, which were customized in a laser cutter (Epilog Legend 36EXT, Epilog Lased, Golden, CO). A simple teeth-box design was constructed in an online box generator MakerCase (https://en.makercase.com) and edited in an open-source graphic editing software Inkscape (http://www.inkscape.org) for the primary structure of the pump. The design included large opening in the walls for easier hand access and a baseplate for more stability (see Supplementary Fig. S1). This design allowed quick assembly and disassembly of the syringe holder that may facilitate portable in situ operation. The dimension of the syringe holder can be adjusted depending on the required application. The design principles for the syringe holder are provided in the Supplementary Fig. S2.

Once laser-cut and assembled, the acrylic walls were held together with tape. Laser-cut, acrylic bars were also taped within the structure for both designs to support the syringe plates within the structure so that they may be swapped out easily. Low-friction acrylic tape (TapeCase Ltd, Elk Grove Village, IL) was applied to the sliding gaps on the pump to reduce the friction between the mass plate and the syringe holder’s walls. Other acrylic components, such as the mass and syringe plates, were also digitally drawn and laser-cut. For each syringe used, its outer diameter was measured with a caliper. This diameter was integrated into Inkscape, so that a simple syringe plate could be fabricated to hold a specific syringe size in place within the system.

Measurement and analysis of pump performance

Frugal microfluidic pump characterization.

To characterize the pump, the pressure exerted by each syringe through loading mass placement on the pump was measured. Plastics syringes (1, 3, 5 mL) (Becton, Dickinson and Company, Franklin Lakes, NJ) and glass syringes (1, 2.5, 5 mL) (Hamilton Company, Reno, NV) were used for characterization.

The setup for pressure characterization involved preparing the syringe pump for our syringe of interest. Each syringe used for characterization was prefilled with DI water. A 30G needle (NE-300PL, Component Supply, Sparta, TN) was connected to the syringe, and microbore tubing (TND80-010, Component Supply, Sparta, TN) was connected to the needle tip. The setup to measure pressure included a digital pressure sensor (Stork Solutions Ltd, Hampshire, UK). Luer lock to ¼ inch NPT adapter (Cole-Palmer, Vernon Hills, IL) was connected to the sensor. Teflon tape was used on the adapter to prevent pressure leakage. The other end of the microbore tubing was then connected to a needle tip that was connected to the adapter. The pressure sensor was connected to a laptop to measure and log pressure data in real time for each experiment.

Before each experiment, care was taken to remove any air bubble in the measuring apparatus. Once connected, the pressure was allowed to stabilize. Preparing for a new trial involved disconnecting the tubing from the pressure sensor, replacing the plastic syringe/emptying and refilling the glass syringe with water, and reconnecting every component back together.

Pressure data was collected in real time through the default software provided by the pressure sensor. Specifically, for syringe characterization, data points were collected after the pressure stabilized from each mass input applied to the system.

Microfluidic device fabrication.

Microfluidic devices composed of polydimethylsiloxane (PDMS) (Sylgard 184 Silicone Elastomer, Dow, Midland, MI) were fabricated from SU-8 molds on silicon wafer (El-Cat Inc., Ridgefield Park, NJ). SU-8 2010 was spin coated on silicon wafer, pattern transferred via a mask exposure (MJB4, KARL SUSS, Germany), and developed using SU-8 developer. After development, the molds were hard baked for 15 minutes at 100°C. To prepare for fabrication of PDMS microfluidic devices, the SU-8 molds were cleansed two times with isopropanol solution and then air dried. These wafers were then placed in a vacuum chamber with 5 μL of Trichloro (1H,1H,2H,2H-perfluorooctyl) silane (Sigma Aldrich, St. Louis, MO) for 1 hour. During this timeframe, PDMS was prepared with a 1:10 ratio of curing agent and PDMS monomer. This mixture was then incorporated into a vacuum chamber for 30 minutes to remove any air bubbles. Aluminum foil was used to create a flat base to hold the treated silicon wafer and PDMS mixture. Once the PDMS was vacuumed and the wafer was fully treated, the wafer was inserted into the aluminum foil base and the PDMS was then poured into it. This was all placed into a vacuum chamber to remove any excess bubbles from the mold + PDMS unit. If there were little-to-no bubbles left in the PDMS, the system was placed into an oven at 70°C for 1 hour to completely cure the PDMS. Once fully cured, the unit was removed from the oven and the aluminum foil was removed entirely. The PDMS was carefully separated from the mold.

For microfluidic device assembly, PDMS was cut into rectangular device segments and punched with the appropriate hole sizes for tubing. Glass slides and PDMS device were cleaned with isopropanol and then scotch tape was used to remove any debris. A plasma wand (BD-10AS High Frequency Generator, Electro Technic Products, Chicago, IL) was used to treat the bonding surfaces of the glass slide and PDMS for 2 minutes. Both components were bonded together and placed in the oven for 1 hour.

Flowrate measurements.

To measure flowrate, fluorescein dye (F6377, Sigma Aldrich, St. Louis, MO) was mixed with DI water. This mixture was placed in a 1 mL plastic syringes and was connected to the microfluidic channel through simple tubing connections. This microfluidic and pump apparatus were placed under an inverted microscope (Nikon Eclipse Ti2 Inverted Microscope, Melville, NY, USA) with a green fluorescent filter. To begin with each experiment, different pressures were applied (50, 75, 100 kPa) to the system. Once the fluorescent dye was close to entering the microfluidic device, a digital CMOS camera (Hamamatsu C11450 ORCA Flash-4.0LT, Bridgewater, NJ, USA), integrated into the microscope, was used to record a time-lapse video. Flowrate was determined by measuring the velocity of the dye in the channel while accounting for the channel geometry.

Different mixtures of DI water, glycerol, and fluorescein dye were utilized to measure flowrates at varying viscosities. The viscosities of each glycerol + DI water solution were prepared such that the ratios between the solution and pure DI water (μ_solution/μ_water) could be set as 1, 5, 10, 20, and 30.

Droplet generation.

Droplet microfluidic devices were fabricated from polydimethylsiloxane (PDMS) and silicon wafer molds. To generate water-in-oil droplets, silicon oil was used as the fluid in continuous phase channel with the water solution as the dispersed phase fluid. This water solution was composed of 0.1% Tween 20 in DI water. Two frugal microfluidic pumps were utilized to control each fluid. Droplet formation was recorded using the high-speed camera. Droplet size was measured from the video footage using Fiji.²⁶

Results and discussion

Characterization of frugal microfluidic pump

This simple pumping system, comprising of only a loading mass and syringe, is able to generate a consistent pressure profile. All syringes tested here achieved a linear pressure response for various mass inputs, as shown in Fig. 2a. Currently, the controllable pressure range spanned from 10 kPa up to 700 kPa. Plastic syringes started to show sufficient exerted pressure at relatively lower loading mass input compared to glass syringes. In particular, 1 mL plastic syringe immediately responds to the smallest loading mass input of 100 gr. On the other hand, 1 mL glass syringe began responding to the loading mass input at 150 gr. Larger syringes require more loading mass input up to a point where considerable exerted pressure can be detected. 3 mL and 5 mL plastic syringes show appreciable pressure at 250 gr, 2.5 mL and 5 mL glass syringes started to show an increase on the exerted pressure at around 350 gr and 600 gr respectively. These differences of initial pressure responses likely occur due to the static friction forces between the syringe and the plunger materials.

Figure 2.

Characterization of the frugal microfluidic pump system. (a) Exerted operational pressure as function of loading mass for different types of syringes. (b) Slope comparison between experimental measurements and theoretical calculation. Left-side columns slopes are plastic syringes and right-side columns are glass syringes slopes. Error bars represent mean ± standard deviation from six independent experiments.

We quantified the loading mass (input) – exerted pressure (output) relationship from the linear slope. The theoretical slope value is derived from Equation 1, in which:

$$Slope(kPa∕gr)=\frac{P_{exerted}}{m_{load}}=\frac{g}{\frac{1}{4}\times \pi \times {d_{syringe}}^2}$$ (2)

We examined the validity of this simple equation to estimate the real exerted pressure from the pump. We calculated the slope from the inner diameter measurements of the syringes listed in Supplementary Table S1. Fig. 2b plots the comparison between theoretical estimates and experimental measurements of the linear slope. There was a small deviation between the theoretical calculation of the slope value and the experimental measurements. This deviation might come from the friction between the rubber plunger and the barrel wall. However, the trend from the theoretical calculation still provided a reliable estimate of the exerted pressure for various types of syringes.

One remarkable finding from our test is that the frugal microfluidic pump can generate precise exerted pressure. The data points in Fig. 2b show measurements from different syringes each time. Despite using a new set of syringes for each of the measurements, the measured slopes were consistent. With this benefit of reproducibility, even with using plastic syringes that are relatively inexpensive, the frugal microfluidic pump can be relied on for many bench-top applications.

A stable pressure for a long period of time is useful for longitudinal assays in microfluidics. We tested the capability of the frugal microfluidic pump to generate a constant pressure for over 12 hours using the 1 mL glass syringe. In Fig. 3a, we show the exerted pressure of the system for various pressures ranging from 50 kPa to 550 kPa. Over 12 hours, for all the pressures tested here, the exerted pressures are constant with less than 5% deviation. Practically, there was no noticeable pressure drop throughout the duration of the experiment.

Figure 3.

Performance of the frugal microfluidic pump system. (a) Long-term pressure characterization. Symbols and error bars represent mean and ± standard deviation from three independent experiments. (b) Dynamic pressure changes characterization. Each data point was recorded at 1 second time resolution. Inset shows the higher temporal resolution plot at the beginning of the experiment.

As described earlier, the choice of syringes for the frugal microfluidic pump needs to be adjusted depending on the required applications. For applications requiring high pressure (over 400 kPa), a 1 mL glass syringe is recommended. Meanwhile, a 1 mL plastic syringe is sufficient to generate moderate pressure (up to 400 kPa), without losing pressure, for an extended period. In some cases, where a large volume of fluids needs to be flown into the microfluidic device, a larger syringe size can be beneficial for this type of utilization.

During microfluidic device operation, some biological assays may require active control over the actuation pressure from the pump. Some of these applications include adjustment of a biochemical signal in a microfluidic perfusion of cell culture chamber,^27-29 observation of the dynamics of cell morphology,³⁰ or optimization of shear stress in microfluidics.³¹ Fig. 3b shows the capability of the frugal microfluidic pump to precisely control the exerted pressure while the device is in operation. When the loading mass input increased, the pump responded immediately in seconds by showing a rapid fine-tuning of the exerted pressure. One limitation of this setup is the residual pressure that is left after decreasing the loading mass input. The pump took minutes to decrease in pressure and stabilize. This finding is similar to the hysteresis phenomenon, where there is a lag of output response between increasing input and decreasing input. Considering this observation, when designing a multistep experiment, it is important to run the pump in ascending order of exerted pressure, from lower pressure to higher pressure.

Flow control in microchannel

Downscaling the microchannel geometry will massively increase its hydrodynamic resistance.³² In planar microfluidics, there is an inverse cubic relationship between the channel height and the hydrodynamic resistance. In other words, the amount of exerted pressure grows rapidly as the microchannel gets smaller. We demonstrated the capability of the frugal microfluidic pump to flow fluid in microfluidic channel with a channel height of 10 μm (Fig. 4a). Following Poiseuille flow, the flow rate is linearly related to the pressure gradient.

Figure 4.

Control of fluid flow in microfluidics using the frugal microfluidic pump. (a) Effect of exerted pressure on flow rate in a microfluidic channel. (b) Flow rate of various viscous fluids in a microfluidic channel. Exerted pressure is 75 kPa. Symbols and error bars represent mean and ± standard deviation from three independent experiments.

The use of viscous fluids also increases the hydrodynamics resistance of the fluid flow. As previously reported, the limited inflation pressure from the latex balloon microfluidic pump made flowing viscous liquid sample challenging.¹⁹ Therefore, we used the frugal microfluidic pump to flow viscous fluid in a small microchannel. Fig. 4b shows that the pumping system is able to successfully drive various viscous fluids into the microfluidic device.

High hydrodynamics resistance when flowing fluid into microfluidics stems from two main factors: microchannel geometry and fluid viscosity. When the channel geometry is small or viscous fluid is being used, flowing fluid into the microfluidic requires an enormous driving pressure at the inlet. This requirement can be unmanageable for other low-cost microfluidic pump systems that have limited working pressure.^19,20 We have demonstrated that the frugal microfluidic pump can generate high pressure to overcome this barrier, lending the system beneficial for microfluidic devices with high hydrodynamics resistance.

Control of droplet size in microfluidic droplet generator

Water-in-oil droplet emulsion has been useful for progressing the field of single cell analysis.³³ In addition to the widely known application in genome sequencing, new developments have shown the utilities of droplet microfluidics for probing cancer cells metabolic rate,³⁴ screening T cell receptor,³⁵ detecting and amplifying micro-RNA,³⁶ and analysing single-cell multi-omics.³⁷ Fig. 5a shows that the frugal microfluidic pump enabled droplet size control in a microfluidic T-junction device. The generated water-in-oil droplet was uniform in size. Increasing the size of the droplet was achieved by reducing the pressure of the continuous phase (oil) while keeping a constant pressure in the dispersed phase (water) (Fig. 5b)

Figure 5.

Microfluidic droplet generation driven by the frugal microfluidic pump. (a) Micrograph of the droplet generator at the T-junction. (b) Effect of continuous phase (oil) inlet pressure on droplet size formation in the microfluidic. Symbols and error bars represent mean and ± standard deviation.

Conclusions

The miniaturization offered by microfluidic and lab-on-chip technology can benefit healthcare infrastructure in low-resource settings. Moreover, microfluidic device has emerged as an invaluable tool in chemistry and biomedicine. As a necessary tool to operate microfluidics, commercial microfluidic pumps are often cost-prohibitive and not field-operable. Despite the increasing demands of microfluidics in many sectors, current instruments to control fluid flow in microfluidics, that are simple, inexpensive, and versatile, are still limited in its output ability and do not provide a single, well-rounded system for many applications. There are open-source microfluidics pumps that use electronics and require experience with instrumentation. However, many microfluidic technologies are used by biologists and chemists who may have limited expertise in instrument development.

To increase the accessibility of microfluidic technology, we have demonstrated a frugal microfluidic pump setup that is simple to fabricate, inexpensive, and controllable. The pump can exert operational pressure of more than 700 kPa (100 Psi), thus lending itself useful for smaller microchannels or even nanofluidic devices. The pressure from the frugal microfluidics was stable for 12 hours, and can possibly be extended to a much longer period of time, which opens up the possibility for longitudinal studies of sample in microfluidics. Moreover, the pressure is controllable during operation, enabling easy parameter optimization of newly developed microfluidics assays. Essentially, the pressure increase of each loading mass input can be tuned to what is required for each application. A high-pressure requirement can utilize smaller syringe size while a low-pressure requirement can benefit from a larger syringe size.

The development and implementation of microfluidics are often not applicable in resource-limited settings in which access for rapid prototyping and device testing is severely limited. The frugal microfluidic pump is expected to propel microfluidic research globally, facilitating the development of appropriate biomedical technology that is more suitable for the developing regions. New biomedical invention intended for use in the resource-limited community is not only indispensable, but also beneficial to improve health care for everyone. With all its versatility, the frugal microfluidic pump will support all stages of microfluidics device life cycle, from development to deployment.